IRAS Explanatory Supplement
XI. Known Processing Anomalies
K. Minor Problems

1. An unmodelled source of position errors was present in the data, as evidenced by the need to increase in-scan (X) for sources that had a cross-scan position error without a significant uniform component.
2. Radiation hits and noise spikes caused some sources to have significantly larger cross-scan errors than quoted.
3. In order to improve completeness, it was necessary to allow detection without a confirming partner due to a failed detector to bypass seconds-confirmation. This resulted in a large number of spurious detections being accorded a status equal to that of a truly seconds-confirming source. These false detections could at times replace valid detections in band-merging.
4. A second effect of failed detections was that an edge detection pair opposite a fulled detector was often not seconds-confirming allowing two separate detections to be present at band-merging. The weaker detection, with its larger error basis, was usually chosen in band-merging resulting in a depressed flux in that band.
5. A third effect of failed detectors resulted from detectors that were impaired but not dead. They would produce detections that were too weak to seconds-confirming with detections on the partner detectors. Again, the weaker detection would often be chosen in band-merging resulting in a depressed flux.
6. Strong radiation hits could also result in a seconds-confirming failure which, in a small fration of cases caused incorrect fluxes or flux status.
7. Responsivity changes due to particle or photon radiation caused baseline changes of order 10% in some observations.
8. The optimum thresholds for accepting the seconds-confirmed sources were set by varying the threshold level and evaluating the numbers of sources passed as a function of threshold value (Section VII.E.5). The method worked satisfactorily in all the wavelength bands but the 25 µm band where the total number never reached the predicted plateau.
9. The CIRR2 flag is incorrectly set to 0, implying no all sky data, for about 40 sources at the 0 = 360' ecliptic longitude boundary.
10. For many catalogs, the truncation instead of rounding of optical magnitudes leads to a 0.1 mag error in the reported magnitudes.
11. For the ESO/Uppsala Survey of the ESO (B) Atlas, only the first two characters of the catalog description were used for the type field.
12. Objects near the ecliptic poles which were scanned twelve times within six successive SOPs produced three separate HCONs. The first and last HCONs were allowed to weeks-confirm normally, while the second HCON was only combined with the source by the clean-up processor. Due to a software design error this program created a small number of sources having two HCONs with identical fluxes.